Israel's AI Infrastructure Wave: Two Rounds, One Signal

<!-- Note: the brief described Edelbyte's round as a "$70M Series A". The company's announcement, reported by Forbes and carried by Morningstar, is $70 million in combined seed and Series A funding. The article below uses the corrected figure. No topic swap was needed. -->

On November 11, 2026, two Israeli AI infrastructure startups announced funding rounds on the same day: Edelbyte raised $70 million and Plurai raised $13 million. Different stages, different products, different backers — but the same underlying bet. The money in the AI economy is moving from the models themselves to the infrastructure that makes agents work: the data layer underneath them and the testing layer that keeps them reliable.

Edelbyte emerged from stealth with $70 million in combined seed and Series A funding, led by Sequoia and Quiet Capital, with participation from Greenoaks, Swish Ventures and Startpoint Capital. The company was founded in 2025 by four data infrastructure veterans — chief executive Avner Aharoni, chief technology officer Evgeney Ryzhyk, Alexander Sloutsky and Yoni Leibowitz — each with about two decades at Microsoft, where they spent more than a decade together building Kusto, the data systems behind dozens of Microsoft products. Edelbyte is building a unified search and analytics platform designed for AI-driven demand, positioned against Snowflake, Databricks and ClickHouse in what the company describes as a $300 billion market. It says it is already running petabyte-scale deployments with Fortune 500 customers, employs 40 people across Israel, the United States and Europe, and is based in Tel Aviv and San Francisco. CTech reported a $500 million valuation.

The Other End of the Stack

Plurai, also founded in 2025, raised a $13 million seed round led by Team8, with participation from U&I, Velocity and Mercer and more than 20 angel investors, including CyberArk founder Udi Mokady and Reah Miyara, senior director of Google Models at Google Cloud and a former model-evaluation lead at OpenAI. Its founders, chief executive Dr. Ilan Kadar and chief technology officer Dr. Elad Levi, previously worked at Nexar, the autonomous-driving company, where Kadar collaborated with Waymo on simulating and evaluating driving software.

Plurai applies the self-driving playbook to AI agents. It builds a digital twin — a software replica — of a company's agent and its operating environment, using the customer's own knowledge base, tools and policies, then simulates thousands of scenarios to find where the agent fails before it reaches production. The platform reached general availability with the funding announcement, adding a free tier for developers after previously being limited to select partners. It works alongside tools such as Claude Code, Codex, Copilot and Cursor. The company employs about 12 people in Israel and the United States and plans to double its workforce by the end of 2026. Longer term, it plans to use simulation data to post-train small, open-weight models tailored to individual customers. Dealroom notes that firms in AI-agent evaluation and observability raised $272 million through September 2026, up from $166 million a year earlier.

Why the Same Day Matters

Two rounds on one day do not make a trend by themselves. But Edelbyte and Plurai sit at opposite ends of the same stack, and together they describe where the agent economy's bottlenecks are. Edelbyte attacks the data problem: agents generate far more queries than human analysts, and the existing analytics platforms were not built for that volume or pattern. Plurai attacks the reliability problem: agents that work in a demo routinely fail in production, and the only proven way to catch those failures is simulation before deployment — the method that got self-driving cars onto roads.

Both companies also share a trait investors are paying a premium for in Israel: founders who have already built the underlying technology once, at scale, inside a large company. The Edelbyte team built Kusto at Microsoft. The Plurai team built simulation systems at Nexar and Waymo. In infrastructure, where buyers need proof the system survives real load, that pedigree is not a nice-to-have; it is the pitch.

The geography matters too. Israel's startup funding tracker recorded both rounds on the same day, and the two companies anchor a Tel Aviv cluster in AI infrastructure that now spans data platforms, agent testing, cybersecurity and developer tools. For a small ecosystem, two significant rounds in one day is a visible signal to global investors that the pipeline is active.

What Founders and Investors Should Take From This

For founders, the message is that infrastructure bets still require real deployments. Edelbyte's claim of petabyte-scale Fortune 500 deployments and Plurai's case studies with NVIDIA and Nebius are what make the rounds credible — not the technology description alone. A new infrastructure startup should assume investors will ask who is running it in production, and at what scale, before they ask about the architecture.

For investors, the two rounds illustrate a maturing agent market. The first wave of agent funding went to the agents themselves. The second wave is going to the plumbing: data, testing, observability and evaluation. That is usually where durable companies are built, because every agent company becomes a customer of the plumbing eventually.

The Israeli AI infrastructure wave is not a single story. It is two stories — data and reliability — arriving on the same day, from the same ecosystem, pointing at the same conclusion: the agent economy is now being built for production, not for demos.

Angeled.com connects founders with investors.